Breaking bad habits: why investors aren’t good at asset allocation

Institutional investors act like momentum investors, chasing returns, even over longer time horizons according to Asset Allocation and Bad Habits, a new research paper that looks at the impact of past returns on asset allocation.

The paper commissioned by Rotman-ICPM and authored by Amit Goyal professor at Univeriste de Lausanne, Andrew Ang professor at Columbia Business School and Antti Ilmanen from AQR Capital Management, empirically documents that longer-horizon investors act like momentum investors.

While many large pension funds rebalance there are also many that let their asset allocation drift with relative asset class performance. This might reflect passive buy and hold policies or a desire to maintain asset allocation near to market cap weights but it can also represent more pro-active return chasing. The paper gives evidence to the latter, using data from CEM Benchmarking on evolving US pension funds’ asset allocations from 1990-2011. It shows return- chasing behaviour at asset class level over multi-year horizons.

One of the authors, Amit Goyal, says investors can be narrow-minded in their decisions around asset allocation.

“They are myopic in this behaviour, they don’t look at asset class returns over a long horizon or even over five years, but more like one year,” he says. “We know from empirical research that returns reverse over three to five years, failure to take that into account is detrimental.”

Goyal says that investors should be considering forward looking economic forecasts in their asset allocation decisions and put less weight on past returns.

“If you are going to make a decision on asset allocation then you need some forecast of future expected returns and risks. But it is like looking into a crystal ball that one doesn’t have. In forming estimates of the future maybe there should be more focus on economic factors and an investor’s own special situation rather than blindly focusing on past returns. Past returns are over-emphasised.”

The research used data from 573 US pension funds which had a median size of $3 billion and an average of around $10 billion. Collectively, the funds hold 30-40 per cent of the assets of US pension funds and about 4 per cent of US equity market capitalisations. The research looked at the funds actual and policy asset allocation weights.

For the period 1990-2011 the policy or strategic target asset allocations, averaged across all funds (equally-weighted) was 57 per cent for equities, 32 per cent for fixed income, 9 per cent for alternatives and 2 per cent for cash.

The analysis shows that policy weights for equities rose from 54 to 61 per cent peak in 1999-2001 before falling to 46 per cent in 2011. Fixed-income weights fell from a third to 29 per cent in 2004-2006 before rising to 35 per cent in 2011, and cash weights had a similar U-shaped time profile. Alternatives weights fell from 10 to 6 per cent in late 1990s before rising to 16 per cent in 2011.

The asset allocation of the funds is analysed alongside momentum/reversal patterns in financial markets.

The paper finds that: “Pension funds in the aggregate do not recognise the shift from momentum to reversal tendencies in asset returns beyond one-year horizon. Pension fund keeps chasing returns over multi-year horizons, to the detriment of the institutions long-run wealth.”

The authors’ hope is that by contrasting the evidence of multi-year pro-cyclical institutional allocations with the findings of multi-year return reversals in many financial assets that it will make at least some investors remedy their bad habits and reconsider their asset allocation practices.”